Anthropic Accuses Chinese AI Firms of Persistent Distillation Attacks

**Anthropic Accuses Chinese AI Firms of Persistent Distillation Attacks Amid Intensified Competition** San Francisco, CA – Anthropic, a leading artificial intelligence research company, has unveiled a new report detailing what it describes as persistent "distillation attacks" carried out by several China-based AI firms. The report, released on Thursday, specifically names prominent players such as Alibaba, Moonshot AI, and DeepSeek as being involved in these campaigns. According to Anthropic, these attacks have seen a significant escalation in recent months, a trend attributed to the increasingly fierce competition within the global AI landscape. Distillation, in the context of AI, refers to a technique where a smaller, simpler model (the student) is trained to replicate the performance and knowledge of a larger, more complex model (the teacher). While model distillation can be a legitimate technique for efficiency, when performed without authorization on proprietary models, it constitutes a form of intellectual property infringement, effectively siphoning off the costly research and development efforts of the original creators. The allegations raise serious concerns about the integrity of AI development and the protection of intellectual property in a rapidly evolving field. As companies invest billions into developing advanced AI models, the unauthorized extraction of their capabilities poses a significant threat to innovation and fair market practices. Anthropic's report underscores a growing tension in the AI sector, where the race to develop superior models is leading to increasingly aggressive tactics. The company has not yet detailed the specific evidence supporting its claims or the exact methods used in these alleged distillation campaigns, but the report signals a more assertive stance against what it perceives as unfair competitive practices.